Young adults (age 21-34) have the highest rates of at-risk drinking and drinking/driving of any age group. Using Problem Behavior Theory as the main conceptual approach, this study will advance theory explaining the development of these behaviors, providing in-depth understanding of their increasing, decreasing, stable, or variable patterns over time. The study will address three aims. [unreadable] [unreadable] Aim 1) Characterize longitudinal young adult patterns of at-risk drinking and drinking/driving. Three young adult surveys will enable longitudinal examination of developmental patterns of risk factors and behaviors during young adulthood. Both linear and non-linear patterns of behavior change/stability over time can be examined, providing a detailed record of the development of at-risk drinking and drinking/driving and its concomitant young adult transitional factors. [unreadable] [unreadable] Aim 2) Determine adolescent characteristics that predict young adult at-risk drinking and drinking/driving. Developmental patterns of risk factors and behaviors during young adulthood can be linked to their precursors in adolescence. Both linear and non-linear adolescent patterns of behavior change/stability over time can be examined, providing a detailed record of the development of at-risk drinking and drinking/driving. Analyses will address an urgently needed understanding of the antecedents, predictors, and concomitants of young adult at-risk drinking and drinking/driving, and can inform the development of future interventions. [unreadable] [unreadable] Aim 3) Develop new statistical models and estimation methods for analyzing combined multiple series of longitudinal data. A unique feature of this study is the availability of subjects' longitudinal data collected from different sources and during different developmental stages. Analyses of such multiple series of longitudinal data require new statistical methodology that will be developed. [unreadable] [unreadable] The unique data set already includes school-based survey data from as early as fifth grade, complete state-reported driver records since subjects were first licensed, and data from a telephone survey conducted when the sample averaged 23 years old. Two additional telephone surveys will be conducted, and driver records will continue to be obtained, yielding over 20 years of rigorously-measured risk factors and behaviors, from early adolescence throughout young adulthood. This project's large, well-characterized, sample allows thorough longitudinal examination of at-risk drinking and drinking/driving. Results from this study will provide the necessary background for developing effective prevention and intervention of these personally and socially costly behaviors.